Login  Email:    Password:   Forget your password? Home      About Us      Products      Uses      Demo      Contact Us

CONCEPTS VS. KEYWORDS: WHAT WE DO

The core element of HTI's semantic offering is the capability to process digital content and automatically and comprehensively identify linguistic concepts expressed within the text. This offering, Skimcast, identifies terms that linguistically behave as if they are carrying meaning within the context of the text, according to the linguistic rules of a language.

This is a novel, powerful, and unique capability of the Skimcast system.

HTI's approach is well beyond statistical key word frequency and co-occurrence in that it automatically identifies and extracts linguistic concepts directly from text itself. Skimcast semantically processes content to provide cleaner, more comprehensive, and more accurate linguistic data in real time that can then be better searched by conventional mathematical, statistical, or user driven search methods.

Once content is processed, the resulting Skimcast output is an exhaustive listing and indexing of all linguistic concepts expressed within the text. This indexing allows readers and/or search engines to search and navigate (and ultimately comprehend) digital content more effectively and more accurately to find the critically important and relevant (and often hard to find or low frequency) content

Skimcast generated concepts can also be placed into larger concept libraries or repositories that allow for infinitely greater and more efficient comprehension and discovery of information expressed across and within all content processed. These Skimcast generated concept libraries provide better search, navigation, and retrieval because, relative to conventional means, they are:

* Exhaustive. Skimcast will automatically identify technical terms including jargon, previously unknown words, and words that may be used in manner inconsistent with their conventional or dictionary meaning (code words). Again, Skimcast concepts are identified based on how they are used within the text. And,

* Clean. Skimcast will only recognize the terms that carry linguistic meaning as defined by the text itself. Skimcast eliminates conventional search "noise" by not recognizing words that do not carry meaning or do not refer to the subject matter, even if the words are used frequently or match a pre-determined search list.

The Skimcast system is built on polynomial complexity and is distributable and linearly scalable. With the proper computing power, Skimcast can process an unlimited volume of digital text.

Skimcast can process content in real time as it does not need to wait for a programmer or analysts to populate a search tool with specific terms or otherwise predict what to search for in content. This makes Skimcast uniquely capable of processing real time content from news feeds, blogs, social media, voice transcripts, and other real time sources.

The system is statistically validated through Ph.D. level research conducted at Cambridge University in England and has been commercially validated by scientific research publishers worldwide. The Skimcast system has also been evaluated by the US Air Force to support proposal analysis, to identify risk information for proposal evaluators.